Airbnb - relation between rent & Airbnb revenue and supply on SA2 level

Data preparation

1 SA2 areas

There are 2,310 SA2 areas in Australia. We excluded areas with no physical location (including 9 Migratory–Offshore–Shipping and 9 No usual address codes for each State and Territory). And 4 remote areas:

##                SA2_NAME16        STE_NAME16
## 1        Lord Howe Island   New South Wales
## 2        Christmas Island Other Territories
## 3 Cocos (Keeling) Islands Other Territories
## 4          Norfolk Island Other Territories

That leaves the initial set of 2,288 areas.

2 APM data

2.1 Raw data

Raw APM dataset consists of 119,656 monthly records of 2,238 SA2 areas for 27 months from 2016-12-01to 2019-02-01.

According to APM

All data in the SA2 file is based on 12-month aggregation.

Therefore the first observation in the dataset for the 2016-12-01 actually represents data from 2016-12-01 to 2016-01-01!

Data are derived separately for houses and units. Selected rent related variables include:

## Observations: 119,656
## Variables: 7
## $ SA2_MAIN16    <dbl> 101021007, 101021007, 101021007, 101021007, 101021007...
## $ date          <date> 2016-12-01, 2017-01-01, 2017-02-01, 2017-03-01, 2017...
## $ record_count  <dbl> 45, 45, 48, 45, 48, 47, 43, 47, 47, 49, 47, 47, 46, 4...
## $ property_type <chr> "House", "House", "House", "House", "House", "House",...
## $ rent_50p      <dbl> 330.0, 330.0, 330.0, 330.0, 330.0, 330.0, 335.0, 335....
## $ rent_25p      <dbl> 290, 290, 290, 290, 290, 295, 300, 300, 300, 300, 300...
## $ rent_75p      <dbl> 365, 360, 365, 365, 365, 370, 375, 380, 380, 375, 380...

2.2 SA2 areas

There are 51 SA2 areas that do not exist in APM data at all

Most seem legit as they mainly consists of natural or military areas, airports, etc. They were excluded from further analyses.

2.3 Missing data

Not all SA2s have information for all months. 21,292 of the records (18%) have missing information on monthly median rent.

2.3.1 Completely missing information

2.3.1.1 Houses

There are 123 SA2 areas that do not have any APM data for houses

These areas are excluded.

2.3.1.2 Units

There are 519 SA2 areas that do not have any APM data for units

These areas are excluded.

2.3.2 Partly missing information

2.3.2.1 Houses

There are 24 SA2 areas that have missing 12 or more months of APM data or have 6 or more consecutive months of APM data missing for houses

These areas are excluded.

2.3.2.2 Units

There are 187 SA2 areas that have missing 12 or more months of APM data or have 6 or more consecutive months of APM data missing for units

These areas are excluded.

2.4 Imputing time series

2.4.1 Houses

There are 24 SA2 areas that have some missing APM data for houses

These areas have data imputed.

2.4.2 Units

There are 94 SA2 areas that have some missing APM data for units

These areas have data imputed.

2.4.3 Example

Data imputed using Stineman interpolation using na_interpolation function from imputeTS package.

2.5 Clean data

Clean APM dataset consists of 56,430 monthly records of house rents in2,090 SA2 areas and 41,337 monthly records of unit rents in1,531 SA2 areas.

2.5.1 Houses

State on 2019-02-01

2.5.2 Units

State on 2019-02-01

3 ABS data

3.1 Selection

Stocks of houses and units were derived from 2016 Census for all areas with APM data.

These were created counting dwellings that were private (Occupied private dwellings and Unoccupied private dwellings from the DWTD Dwelling Type variable) and then divided into houses and apratments uing STRD Dwelling Structure variable with:

• Separate house, Semi-detached, row or terrace house, townhouse etc. with one storey and Semi-detached, row or terrace house, townhouse etc. with two or more storeys representing houses

• Flat or apartment in a one or two storey block, Flat or apartment in a three storey block, Flat or apartment in a four or more storey block, Flat or apartment attached to a house representing units

3.2 Results

3.3 No ABS units

Note: there are 25 SA2s with zero ABS units, but available APM rental units!

SA2_NAME16	min	mean	max
Bateman	14	20.259259	24
Beaconsfield - Officer	9	14.444444	20
Brassall	112	128.666667	159
Calwell	7	13.888889	19
College Grove - Carey Park	44	57.407407	66
Coober Pedy	9	13.481481	16
Craigie - Beldon	14	17.296296	20
Dingley Village	11	16.111111	19
Edens Landing - Holmview	28	35.888889	49
Endeavour Hills - South	14	18.888889	25
Lawson	6	49.518518	78
Leeming	11	14.111111	17
Loganholme - Tanah Merah	20	24.666667	28
Marangaroo	8	10.481481	13
Mount Dandenong - Olinda	6	9.962963	14
Murrumba Downs - Griffin	144	178.888889	250
Narangba	11	19.185185	28
Ngunnawal	40	48.444444	57
Padbury	14	19.333333	24
Quoiba - Spreyton	8	12.000000	16
Reedy Creek - Andrews	21	25.481482	31
Ripley	7	16.148148	24
Rosedale	2	9.444444	13
Seabrook	10	11.851852	16
Wakerley	23	30.222222	40

3.4 Discrepancies

Discrepancies between APM and ABS stocks

SA2_NAME16	property_type	abs_supply	max_record_count	apm_surplus
North Ryde - East Ryde	Unit	364	692	328
Dakabin - Kallangur	Unit	72	368	296
Murrumba Downs - Griffin	Unit	0	250	250
Toowoomba - West	Unit	28	251	223
Brassall	Unit	0	159	159
Lavington	Unit	217	372	155
Eimeo - Rural View	Unit	26	180	154
Albury - North	Unit	32	173	141
Asquith - Mount Colah	Unit	175	314	139
Redbank Plains	Unit	41	180	139
Shepparton - North	Unit	66	200	134
Cambooya - Wyreema	Unit	3	133	130
Rouse Hill - Beaumont Hills	Unit	140	268	128
Upper Coomera - Willow Vale	Unit	8	130	122
Pimpama	Unit	4	118	114
Muswellbrook	Unit	49	158	109
Pacific Pines - Gaven	Unit	4	111	107
Highton	Unit	52	148	96
Werribee - East	Unit	34	130	96
Rangeville	Unit	17	113	96
Tullamarine	Unit	8	102	94
Calamvale - Stretton	Unit	18	111	93
North Toowoomba - Harlaxton	Unit	57	146	89
Busselton	Unit	57	144	87
Flora Hill - Spring Gully	Unit	41	126	85
East Mackay	Unit	19	104	85
Bellbird Park - Brookwater	Unit	21	104	83
Burpengary	Unit	12	94	82
Boronia Heights - Park Ridge	Unit	7	87	80
Caloundra - West	Unit	32	112	80
Viewbank - Yallambie	Unit	14	93	79
Lawson	Unit	0	78	78
Wodonga	Unit	129	205	76
Tamworth - East	Unit	190	265	75
Belmont	Unit	98	172	74
Kellyville	Unit	120	193	73
Ringwood East	Unit	188	259	71
College Grove - Carey Park	Unit	0	66	66
Palmerston - South	Unit	52	118	66
Hastings - Somers	Unit	47	112	65
Strathmore	Unit	27	88	61
Croydon Hills - Warranwood	Unit	6	67	61
Laverton	Unit	43	104	61
Peregian Springs	Unit	22	83	61
Coombabah	Unit	7	67	60
Larapinta	Unit	22	80	58
Newcomb - Moolap	Unit	69	126	57
Ngunnawal	Unit	0	57	57
Epping - South	Unit	24	79	55
Manly West	Unit	20	74	54
Goodna	Unit	21	75	54
Mount Warren Park	Unit	6	59	53
East Bendigo - Kennington	Unit	79	131	52
Jimboomba	Unit	4	56	52
Pakenham - South	Unit	27	78	51
Croydon - West	Unit	41	91	50
Edens Landing - Holmview	Unit	0	49	49
Banjup	Unit	5	54	49
Newnham - Mayfield	Unit	18	67	49
Raymond Terrace	Unit	64	112	48
Colac	Unit	39	86	47
Springfield Lakes	Unit	3	49	46
Salisbury North	Unit	10	56	46
Morphett Vale - East	Unit	39	85	46
Maryland - Fletcher - Minmi	Unit	17	62	45
Melton	Unit	14	59	45
Kingston (Qld.)	Unit	49	94	45
Morayfield - East	Unit	6	51	45
Kirwan - West	Unit	27	72	45
Eaton - Pelican Point	Unit	10	55	45
Wantirna	Unit	3	47	44
Mooroolbark	Unit	29	72	43
Albury - East	Unit	143	184	41
Knoxfield - Scoresby	Unit	10	51	41
Albany	Unit	65	106	41
Wakerley	Unit	0	40	40
Cashmere	Unit	3	43	40
Clarkson	Unit	30	70	40
Hallam	Unit	3	42	39
Leichhardt - One Mile	Unit	37	75	38
Woodridge	Unit	151	189	38
The Hills District	Unit	28	66	38
Windsor - Bligh Park	Unit	77	114	37
Somerville	Unit	3	40	37
Raceview	Unit	42	79	37
Highfields	Unit	4	41	37
Geraldton - East	Unit	14	51	37
Latrobe	Unit	6	43	37
Riverstone - Marsden Park	Unit	14	50	36
Collinsville	Unit	10	46	36
Deeragun	Unit	27	63	36
Bonython	Unit	14	50	36
Murarrie	Unit	34	69	35
Crestmead	Unit	3	38	35
Palmerston	Unit	4	39	35
Runcorn	Unit	52	86	34
Ooralea - Bakers Creek	Unit	4	38	34
Bald Hills	Unit	3	36	33
Aitkenvale	Unit	69	102	33
Sunnybank	Unit	21	53	32
Cobbitty - Leppington	Unit	17	48	31
Norman Gardens	Unit	117	148	31
Reedy Creek - Andrews	Unit	0	31	31
Mount Barker	Unit	50	81	31
Balcatta - Hamersley	Unit	29	60	31
Monash	Unit	3	34	31
Wynyard	Unit	7	37	30
Airport West	Unit	67	96	29
Gisborne	Unit	4	33	29
Riverton - Shelley - Rossmoyne	Unit	50	79	29
Boulder	Unit	26	55	29
Hampton Park - Lynbrook	Unit	48	76	28
Mooroopna	Unit	37	65	28
Park Avenue	Unit	38	66	28
Loganholme - Tanah Merah	Unit	0	28	28
Narangba	Unit	0	28	28
Wyndham Vale	Unit	20	47	27
Kalamunda - Maida Vale - Gooseberry Hill	Unit	5	32	27
Carina Heights	Unit	127	153	26
Kalgoorlie - North	Unit	38	64	26
Endeavour Hills - South	Unit	0	25	25
Bundoora - West	Unit	37	61	24
Coomera	Unit	155	179	24
Ripley	Unit	0	24	24
Padbury	Unit	0	24	24
Bateman	Unit	0	24	24
Warnervale - Wadalba	Unit	3	26	23
Churchill - Yamanto	Unit	5	27	22
Morphett Vale - West	Unit	18	40	22
Duncraig	Unit	8	30	22
Greenwood - Warwick	Unit	4	26	22
Amaroo	Unit	5	27	22
Nicholls	Unit	6	28	22
Truganina	Unit	29	50	21
Browns Plains	Unit	20	41	21
Ross	Unit	7	28	21
Lara	Unit	37	57	20
Beaconsfield - Officer	Unit	0	20	20
Craigie - Beldon	Unit	0	20	20
Girrawheen	Unit	11	31	20
Coogee	Unit	3	23	20
Mount Annan - Currans Hill	Unit	5	24	19
South Morang (North)	Unit	3	22	19
Dingley Village	Unit	0	19	19
Marsden	Unit	84	103	19
Heatley	Unit	23	42	19
Morley	Unit	49	68	19
Calwell	Unit	0	19	19
Wagga Wagga - North	Unit	21	39	18
Ferntree Gully (South) - Upper Ferntree Gully	Unit	34	52	18
Bundaberg Region - North	Unit	12	30	18
Margaret River	Unit	21	39	18
Woodend	Unit	9	26	17
Parkinson - Drewvale	Unit	3	20	17
Kelso	Unit	14	31	17
Leeming	Unit	0	17	17
Meadow Heights	Unit	50	66	16
Seabrook	Unit	0	16	16
Melton West	Unit	30	46	16
Bli Bli	Unit	15	31	16
Nairne	Unit	3	19	16
Golden Grove	Unit	3	19	16
Coober Pedy	Unit	0	16	16
Forrestfield - Wattle Grove	Unit	4	20	16
Sorell - Richmond	Unit	15	31	16
Quoiba - Spreyton	Unit	0	16	16
Watsonia	Unit	19	34	15
Mill Park - South	Unit	22	37	15
Mulgrave	Unit	41	56	15
Busselton Region	Unit	14	29	15
Anna Bay	Unit	8	22	14
Macquarie Fields - Glenfield	Unit	41	55	14
Leopold	Unit	10	24	14
Ringwood North	Unit	12	26	14
Lilydale - Coldstream	Unit	68	82	14
Mount Dandenong - Olinda	Unit	0	14	14
Clarinda - Oakleigh South	Unit	28	42	14
Parkhurst - Kawana	Unit	54	68	14
Geraldton - South	Unit	15	29	14
Claymore - Eagle Vale - Raby	Unit	7	20	13
Rosedale	Unit	0	13	13
Vermont	Unit	39	52	13
Koo Wee Rup	Unit	3	16	13
Svensson Heights - Norville	Unit	49	62	13
Madeley - Darch - Landsdale	Unit	3	16	13
Marangaroo	Unit	0	13	13
Skye - Sandhurst	Unit	4	16	12
Arundel	Unit	41	53	12
Inala - Richlands	Unit	106	118	12
Beerwah	Unit	38	50	12
Pinjarra	Unit	6	18	12
Bonner	Unit	13	25	12
Murray Bridge	Unit	70	81	11
Florey	Unit	16	27	11
Hadfield	Unit	32	42	10
Happy Valley	Unit	3	13	10
Warnbro	Unit	5	15	10
Palmerston - North	Unit	23	33	10
Callala Bay - Currarong	Unit	11	20	9
Nunawading	Unit	147	156	9
Gawler - North	Unit	9	18	9
Green Valley	Unit	9	17	8
Lowood	Unit	33	41	8
Hillcrest	Unit	18	26	8
Kariong	Unit	10	17	7
Niddrie - Essendon West	Unit	81	88	7
Gladstone Park - Westmeadows	Unit	23	30	7
Telina - Toolooa	Unit	51	58	7
Romaine - Havenview	Unit	22	29	7
Maitland - North	Unit	12	18	6
Taree Region	Unit	22	28	6
Carrum Downs	Unit	72	78	6
Alexandra Hills	Unit	49	55	6
Mount Louisa	Unit	10	16	6
Mundingburra	Unit	75	81	6
Jingili	Unit	11	17	6
Condell Park	Unit	51	56	5
Wattle Glen - Diamond Creek	Unit	14	19	5
Keilor East	Unit	95	100	5
Oxenford - Maudsland	Unit	18	23	5
Ipswich - North	Unit	13	18	5
Kirwan - East	Unit	79	84	5
Collie	Unit	22	27	5
Baldivis	Unit	9	14	5
Narrogin	Unit	19	24	5
Delacombe	Unit	15	19	4
Rowville - Central	Unit	51	55	4
Darra - Sumner	Unit	24	28	4
Charles	Unit	91	95	4
Forde	Unit	19	23	4
West Hoxton - Middleton Grange	Unit	19	22	3
Falcon - Wannanup	Unit	20	23	3
Greenfields	Unit	30	33	3
Wanneroo	Unit	21	24	3
Cooloongup	Unit	16	19	3
Kambalda - Coolgardie - Norseman	Unit	58	61	3
Templestowe Lower	Unit	59	61	2
Collingwood Park - Redbank	Unit	37	39	2
Katanning	Unit	25	27	2
Somerset	Unit	18	20	2
Alexandra	Unit	26	27	1
Nuriootpa	Unit	21	22	1
Armadale - Wungong - Brookdale	Unit	172	173	1
Acton - Upper Burnie	Unit	22	23	1
Lawson	House	60	82	22
Melbourne	House	73	82	9

4 Airbnb data

Data prep focusing on monthly income generated on SA2 level.

Raw datasets consist of 359,300 properties and 6,547,439 monthly stats.

4.1 Data preparation

4.1.1 Selection

325,967 data points from before 2016-01-01 were excluded.

101,329 data points of listing_type “Shared room” and 2,304,488 data points of “Private room” were excluded.

480 records of 26 properties located in remote areas were excluded.

Next, 15,189 records of 847 properties located in 104 areas where no APM data is available

Next, 4,123 monthly records of 4,008 properties that had zero AUD income but some information in USD had income recalculated using 1.4 exchange rate.

Next, 3 monthly records of 3 properties that had missing AUD income but some information in USD had income recalculated using 1.4 exchange rate.

4.1.2 Outliers

There are few ouliers in terms of monthly income

## # A tibble: 36 x 7
##    property_id reporting_month reservation_days revenue_usd revenue_native
##          <dbl> <date>                     <dbl>       <dbl>          <dbl>
##  1     6382471 2018-05-01                     3        249          677160
##  2     6382471 2018-04-01                     1         83          225720
##  3    14999741 2018-10-01                    30     148560          209700
##  4    14999741 2018-05-01                    27     146529          188730
##  5    14999741 2019-01-01                    27     136782          188730
##  6    14999741 2018-07-01                    26     134498          181740
##  7    14999741 2018-12-01                    26     128206          181740
##  8    14999741 2019-02-01                    26     129480          181740
##  9     4144830 2016-12-01                    22       2196.         176347
## 10    14999741 2018-08-01                    24     120648          167760
## # ... with 26 more rows, and 2 more variables: adr_usd <dbl>, adr_native <dbl>

From the top monthly income values (with income above $95k) data with very high AUD income were corrected using USD values.

One very high value was replaced to average ADR * booked days from previous & next month from the same property.

Three very high monthly values were recalculated from more realistic ADR (in USD) indicators.

4.1.3 Inactive listings

Over third of monthly records is labelled as not active.

## 
## active <lgl>
## # total N=3941540  valid N=3941540  mean=0.62  sd=0.49
## 
##    val     frq raw.prc valid.prc cum.prc
##   TRUE 2439163   61.88     61.88   61.88
##  FALSE 1502377   38.12     38.12  100.00
##   <NA>       0    0.00        NA      NA

All of them were excluded. That also applied to records where there was some information about the revenue (either AUD or USD; n = 1,128) or number of available days was indicated to be > 0 (n = 448,859). For the former group, that could indicate some reshufling of money from previous months? For the latter group - this doesnt make sense but at least none of these records has any reservations and there are not much money involved.

4.1.4 Missing revenue info

628,370 records of 105,318 properties with missing info on revenue_native were recoded to zero.

property_id	reporting_month	reservation_days	revenue_native
12936	2016-01-01	12	3082.12
12936	2016-02-01	0	NA
12936	2016-03-01	0	NA
12936	2016-04-01	0	NA
12936	2016-05-01	0	NA
12936	2016-06-01	0	NA
12936	2016-07-01	0	NA
12936	2016-08-01	0	NA
12936	2016-09-01	0	NA
12936	2016-10-01	0	NA
12936	2016-11-01	0	NA
12936	2016-12-01	3	671.89

After correction it looks like:

property_id	reporting_month	reservation_days	revenue_native
12936	2016-01-01	12	3082.12
12936	2016-12-01	3	671.89
12936	2017-01-01	5	998.78
12936	2017-02-01	0	0.00
12936	2017-03-01	0	0.00
12936	2017-04-01	0	0.00
12936	2017-05-01	0	0.00
12936	2017-06-01	0	0.00
12936	2017-07-01	0	0.00
12936	2017-12-01	8	1208.80
12936	2018-01-01	14	1864.76
12936	2018-02-01	5	769.49

4.1.5 0 bedrooms

11,214 properties with 0 bedrooms were recoded to have 1 bedroom.

4.1.6 Property type

Analysed properties come in different flavours

## 
## property_type <character>
## # total N=2439163  valid N=2439138  mean=35.49  sd=30.85
## 
##                             val    frq raw.prc valid.prc cum.prc
##                       Apartment 893883   36.65     36.65   36.65
##                           House 876897   35.95     35.95   72.60
##                       Townhouse  83286    3.41      3.41   76.01
##                         Cottage  70485    2.89      2.89   78.90
##                     Guest suite  68681    2.82      2.82   81.72
##                      Guesthouse  61447    2.52      2.52   84.24
##                Entire apartment  48333    1.98      1.98   86.22
##                           Villa  46257    1.90      1.90   88.12
##                    Entire house  42642    1.75      1.75   89.86
##                           Cabin  41546    1.70      1.70   91.57
##                     Condominium  36773    1.51      1.51   93.08
##                        Bungalow  29406    1.21      1.21   94.28
##              Serviced apartment  24898    1.02      1.02   95.30
##                       Farm stay  22305    0.91      0.91   96.22
##                            Loft  14413    0.59      0.59   96.81
##                          Chalet  12044    0.49      0.49   97.30
##               Bed and breakfast  11074    0.45      0.45   97.75
##                           Place   6926    0.28      0.28   98.04
##                       Camper/RV   4739    0.19      0.19   98.23
##                      Tiny house   3368    0.14      0.14   98.37
##                Entire townhouse   3174    0.13      0.13   98.50
##                     Earth house   2511    0.10      0.10   98.60
##                            Boat   2484    0.10      0.10   98.71
##                            Tent   2297    0.09      0.09   98.80
##                           Other   2179    0.09      0.09   98.89
##                   Vacation home   2167    0.09      0.09   98.98
##                    Entire villa   2014    0.08      0.08   99.06
##                 Bed & Breakfast   1855    0.08      0.08   99.14
##                            Barn   1525    0.06      0.06   99.20
##                    Entire place   1451    0.06      0.06   99.26
##                        Campsite   1396    0.06      0.06   99.32
##                    Entire cabin   1349    0.06      0.06   99.37
##            Entire vacation home   1326    0.05      0.05   99.43
##                             Hut   1261    0.05      0.05   99.48
##                       Treehouse   1167    0.05      0.05   99.53
##                           Train   1024    0.04      0.04   99.57
##                 Entire bungalow    823    0.03      0.03   99.60
##               Entire guesthouse    635    0.03      0.03   99.63
##                    Nature lodge    617    0.03      0.03   99.65
##          Entire bed & breakfast    554    0.02      0.02   99.68
##                            Tipi    515    0.02      0.02   99.70
##                            Yurt    490    0.02      0.02   99.72
##              Entire guest suite    472    0.02      0.02   99.74
##                          Island    417    0.02      0.02   99.75
##                   Entire chalet    378    0.02      0.02   99.77
##                          Resort    367    0.02      0.02   99.78
##                          Castle    358    0.01      0.01   99.80
##              Room in aparthotel    340    0.01      0.01   99.81
##                     Entire loft    336    0.01      0.01   99.83
##                      Dome house    304    0.01      0.01   99.84
##          Room in boutique hotel    300    0.01      0.01   99.85
##              Entire condominium    269    0.01      0.01   99.86
##                     Entire tent    260    0.01      0.01   99.87
##                       Houseboat    241    0.01      0.01   99.88
##                Entire camper/RV    218    0.01      0.01   99.89
##       Entire serviced apartment    211    0.01      0.01   99.90
##                           Floor    206    0.01      0.01   99.91
##                            Cave    204    0.01      0.01   99.92
##                       Timeshare    177    0.01      0.01   99.92
##                        Home/apt    169    0.01      0.01   99.93
##                          Hostel    168    0.01      0.01   99.94
##        Entire bed and breakfast    158    0.01      0.01   99.94
##             Bed &amp; Breakfast    148    0.01      0.01   99.95
##                             Bus    101    0.00      0.00   99.95
##                   Room in hotel     98    0.00      0.00   99.96
##                     Entire boat     91    0.00      0.00   99.96
##           Entire boutique hotel     91    0.00      0.00   99.97
##                  Boutique hotel     90    0.00      0.00   99.97
##                           Igloo     80    0.00      0.00   99.97
##                   Entire in-law     65    0.00      0.00   99.98
##             Entire nature lodge     64    0.00      0.00   99.98
##                      Lighthouse     64    0.00      0.00   99.98
##                            Dorm     51    0.00      0.00   99.98
##              Entire earth house     51    0.00      0.00   99.98
##                     Earth House     47    0.00      0.00   99.99
##                  Entire cottage     39    0.00      0.00   99.99
##                     Entire tipi     38    0.00      0.00   99.99
##                     Entire yurt     38    0.00      0.00   99.99
##                    Entire Floor     32    0.00      0.00   99.99
##                 Casa particular     27    0.00      0.00   99.99
##                Entire timeshare     27    0.00      0.00   99.99
##                          In-law     27    0.00      0.00  100.00
##                Entire treehouse     24    0.00      0.00  100.00
##                     Entire dorm     19    0.00      0.00  100.00
##  Room in heritage hotel (india)     18    0.00      0.00  100.00
##                           Plane     15    0.00      0.00  100.00
##                            Flat     11    0.00      0.00  100.00
##                     Entire cave      8    0.00      0.00  100.00
##                      Entire hut      3    0.00      0.00  100.00
##                         Pousada      1    0.00      0.00  100.00
##                            <NA>     25    0.00        NA      NA

Classification of properties was simplified to house and apartment

property_type	property_type_new
Apartment	Unit
House	House
Townhouse	House
Entire apartment	Unit
Bed and breakfast	NA
Guest suite	NA
Entire house	House
Guesthouse	House
Villa	House
Cottage	House
Private room in house	NA
Private room in apartment	NA
Condominium	Unit
Cabin	House
Bungalow	House
Serviced apartment	Unit
Farm stay	NA
Bed & Breakfast	NA
Loft	Unit
Chalet	House
Other	NA
Private room	NA
Room in boutique hotel	NA
Hostel	NA
Place	NA
Private room in townhouse	NA
Entire townhouse	House
Camper/RV	NA
Tent	NA
Bed & Breakfast	NA
Nature lodge	NA
Room in hotel	NA
Boat	NA
Vacation home	House
Private room in bed & breakfast	NA
Shared room in apartment	NA
Tiny house	NA
Entire villa	House
Earth house	NA
Resort	NA
Dorm	NA
Entire place	NA
Entire vacation home	House
Entire cabin	House
Room in aparthotel	NA
Private room in bed and breakfast	NA
Campsite	NA
Shared room in house	NA
Private room in villa	NA
Hut	House
Treehouse	NA
Barn	House
Private room in guest suite	NA
Camper/rv	NA
Entire bungalow	House
Private room in guesthouse	NA
Boutique hotel	NA
Entire guesthouse	House
Tipi	NA
Train	NA
Private room in condominium	NA
Entire guest suite	NA
Entire condominium	Unit
Yurt	NA
Entire bed & breakfast	NA
Castle	House
Shared room	NA
Private room in boutique hotel	NA
Island	NA
Entire chalet	House
Entire serviced apartment	Unit
Entire loft	Unit
Private room in cabin	NA
Entire camper/RV	NA
Private room in hostel	NA
Dome house	House
Private room in bungalow	NA
Floor	NA
Timeshare	NA
Private room in dorm	NA
Entire tent	NA
Shared room in dorm	NA
Private room in serviced apartment	NA
Shared room in townhouse	NA
Earth House	NA
Entire Floor	NA
Shared room in hostel	NA
Private room in loft	NA
Cave	NA
Houseboat	NA
Private room in vacation home	NA
Private room in camper/rv	NA
Entire boat	NA
Shared room in condominium	NA
Entire bed and breakfast	NA
Shared room in villa	NA
Private room in tent	NA
Entire boutique hotel	NA
Home/apt	NA
Casa particular (cuba)	House
Private room in boat	NA
Igloo	NA
Shared room in loft	NA
Private room in nature lodge	NA
Entire timeshare	NA
In-law	NA
Entire in-law	NA
Shared room in guesthouse	NA
Bus	NA
Shared room in bed & breakfast	NA
Entire nature lodge	NA
Private room in chalet	NA
Lighthouse	NA
Entire cottage	House
Entire earth house	NA
Parking Space	NA
Entire tipi	NA
Entire dorm	NA
Entire yurt	NA
Entire treehouse	NA
Private room in hut	NA
Casa particular	House
Hotel	NA
Private room in earth house	NA
Shared room in guest suite	NA
Private room in timeshare	NA
Room in heritage hotel (india)	NA
Private room in cottage	NA
Private room in heritage hotel (india)	NA
Private room in yurt	NA
Shared room in nature lodge	NA
Private room in treehouse	NA
Aparthotel	NA
Private room in in-law	NA
Private room in farm stay	NA
Shared room in bed and breakfast	NA
Shared room in tipi	NA
Shared room in tent	NA
Private room in floor	NA
Plane	NA
Flat	Unit
Shared room in vacation home	NA
Shared room in boutique hotel	NA
Private room in island	NA
Shared room in boat	NA
Shared room in cottage	NA
Van	NA
Entire cave	NA
Private room in parking space	NA
Ryokan (japan)	NA
Private room in tipi	NA
Entire hut	House
Heritage hotel (india)	NA
Minsu (taiwan)	NA
Shared room in camper/rv	NA
Private room in aparthotel	NA
Private room in castle	NA
Private room in tiny house	NA
Pousada	NA
Shared room in cave	NA
Shared room in serviced apartment	NA
Shared room in in-law	NA
Trullo (italy)	NA

140,643 ambiguous/imprecise monthly records of 9,959 properties were excluded

## 
## property_type <character>
## # total N=2298520  valid N=2298520  mean=1.44  sd=0.50
## 
##    val     frq raw.prc valid.prc cum.prc
##  House 1279393   55.66     55.66   55.66
##   Unit 1019127   44.34     44.34  100.00
##   <NA>       0    0.00        NA      NA

4.2 Prepared dataset

Final dataset consists of 2,298,520 monthly records of 202,862 properties in 2,040 SA2 areas.

4.2.1 property_type vs bedrooms

Properties in Airbnb and APM come in different sizes. It is possible to investigate it in the former dataset but not the latter. This might be potential weakness since the rents & ADR must be somehow realted to size.

bedrooms	House	Unit	Total
1	13% (12777)	46% (46259)	29% (59036)
2	22% (22884)	44% (44107)	33% (66991)
3	35% (35285)	9% (9482)	22% (44767)
4	22% (22347)	1% (927)	11% (23274)
5	6% (6422)	0% (121)	3% (6543)
6	1% (1456)	0% (47)	1% (1503)
7	0% (386)	0% (13)	0% (399)
8	0% (142)	0% (3)	0% (145)
9	0% (51)	0% (3)	0% (54)
10	0% (54)	0% (6)	0% (60)
11	0% (7)	0% (0)	0% (7)
12	0% (7)	0% (0)	0% (7)
13	0% (3)	0% (0)	0% (3)
14	0% (7)	0% (1)	0% (8)
15	0% (3)	0% (1)	0% (4)
16	0% (1)	0% (0)	0% (1)
19	0% (1)	0% (0)	0% (1)
20	0% (2)	0% (0)	0% (2)
22	0% (1)	0% (0)	0% (1)
26	0% (1)	0% (0)	0% (1)
30	0% (0)	0% (1)	0% (1)
NA	0% (20)	0% (34)	0% (54)
Total	100% (101857)	100% (101005)	100% (202862)

4.2.2 SA2 areas

Locations of Airbnb properties (points) were linked to ABS data of SA2 areas from 2016. Spatial join in ArcGIS was used with CLOSEST option to capture locations that did not overlap with polygons (see example here https://www.airbnb.com/rooms/19103554 - due to privacy reasons, locations are not exact).

248 SA2s (out of 2288) have no Airbnbs inside.

Few of this areas have some info in APM data.

Houses (n = 53):

Units (n = 25):

Supply & income of Airbnbs inside these areas will be later fixed to zero.

4.2.3 Income of properties

4.2.3.1 Individual revenue

Each property has a ‘trajectory’ of income data. But not all properties have info on all months.. therefore not all SA2s will have income info for all months..

Such coverage issue also applies to ADR.

Note that when income is 0 AUD, ADR is (obviously) missing!

4.2.3.2 SA2 revenue

Individual properties’ ADR with smoothed trend:

Monthly aggregated data with monthly medians of ADR:

4.2.3.3 Averages from last 12 months

4.2.3.4 ‘Filling’ time series

Areas where no income was reported during specific reporting month had those months filled with 0 AUD.

Result is in the form of ‘income curves’ with 97767 data points of Airbnb revenue for 2288 areas.

4.2.3.5 Missing data

Not all SA2s have information for all months. 16,582 of the records (17%) have missing information on monthly median ADR.

4.2.3.5.1 Completely missing information

4.2.3.5.2 Houses

There are 79 SA2 areas that do not have any Airbnb data for houses

These areas are excluded.

4.2.3.5.3 Units

There are 277 SA2 areas that do not have any Airbnb data for units

These areas are excluded.

4.2.3.6 Partly missing information

4.2.3.6.1 Houses

There are 242 SA2 areas that have missing 12 or more months of Airbnb data or have 6 or more consecutive months of Airbnb data missing for houses

These areas are excluded.

4.2.3.6.2 Units

There are 217 SA2 areas that have missing 12 or more months of Airbnb data or have 6 or more consecutive months of Airbnb data missing for units

These areas are excluded.

4.2.3.7 Imputing time series

4.2.3.7.1 Houses

There are 120 SA2 areas that have some missing Airbnb data for houses

These areas have data imputed.

4.2.3.7.2 Units

There are 95 SA2 areas that have some missing Airbnb data for units

These areas have data imputed.

4.2.3.7.3 Example

Data imputed using Stineman interpolation using na_interpolation function from imputeTS package.

4.2.3.8 ADR in SA2

Unit

House

State on 2016-12-01 vs 2019-02-01 houses

State on 2016-12-01 vs 2019-02-01 units

##            used  (Mb) gc trigger   (Mb)  max used   (Mb)
## Ncells  3863322 206.4    7171400  383.0   7171400  383.0
## Vcells 66402005 506.7  187470928 1430.3 593047525 4524.6

4.2.4 Supply & demand of properties

Denomionator for the following two measures was derived from ABS data described above.

Note that this denominator is a ‘static’ variable without temporal variation.

4.2.4.1 Binary

This is naive indicator that takes into account all active properties listed in given month in SA2 irrespectively if they were used or not. Example of one area values

4.2.4.1.1 ‘Filling’ time series and dealing with zeroes

Areas where no properties were listed during specific reporting month had those months filled with 0. Relative measure for each month was calculated using number of properties (numerator) and ABS Census stocks of given property type (denominator).

Areas with zero stocks of listings had this measure fixed to zero.

Areas with zero stocks of census properties and any number of propoerties had this measure fixed to missing.

Result is in the form of ‘supply curves’ with 106628 data points of Airbnb counts for 2288 areas.

4.2.4.1.2 Total properties in SA2

Unit

House

State on 2019-02-01

4.2.4.1.3 Relative properties in SA2

Unit

House

State on 2019-02-01

‘No numbers’ of relative supply, Airbnb supply > 0, ABS stock == 0

‘Large numbers’ relative supply of >= 1, ABS stock > 0

‘Normal numbers’ relative supply of < 1, ABS stock > 0

4.2.4.2 Guest Nights Available (GNA)

This is a bit more elaborate indicator that takes into account all active properties listed in given month in SA2 and uses their available_days as measure.

4.2.4.2.1 ‘Filling’ time series and dealing with zeroes

Areas where no properties were listed during specific reporting month had those months filled with 0. Relative measure for each month was calculated using number of available_days (numerator) and ABS Census stocks of given property type multiplied by number of days in a given month (denominator).

Areas with zero stocks of census properties and zero propoerties had this measure fixed to zero.

Areas with zero stocks of census properties and any number of propoerties had this measure fixed to missing.

Result is in the form of ‘GNO curves’ with 106628 data points of Airbnb counts for 2288 areas.

4.2.4.2.2 Total GNA in SA2

Unit

House

State on 2019-02-01

4.2.4.2.3 Relative GNA in SA2

Unit

House

State on 2019-02-01

‘No numbers’ of relative supply, Airbnb supply > 0 ABS stock == 0

‘Large numbers’ relative supply of >= 1, ABS stock > 0

‘Normal numbers’ relative supply of < 1, ABS stock > 0

4.2.4.3 Guest Nights Occupied (GNO)

This is again bit more elaborate indicator that takes into account all active properties listed in given month in SA2 and uses their reservation_days as measure.

Number of reservation days:

4.2.4.3.1 ‘Filling’ time series and dealing with zeroes

Areas where no properties were listed during specific reporting month had those months filled with 0.

Relative measure for each month was calculated using number of reservation_days (numerator) and ABS Census stocks of given property type multiplied by number of days in a given month (denominator).

Areas with zero stocks of census properties and zero propoerties had this measure fixed to zero.

Areas with zero stocks of census properties and any number of propoerties had this measure fixed to missing.

Result is in the form of ‘GNO curves’ with 106628 data points of Airbnb counts for 2288 areas.

4.2.4.3.2 Total GNO in SA2

Unit

House

State on 2019-02-01

4.2.4.3.3 Relative GNO in SA2

Unit

House

State on 2019-02-01

‘No numbers’ of relative supply, demand_days> 0 & ABS stock == 0

‘Large numbers’ relative supply of >= 1, ABS stock > 0

‘Normal numbers’ relative supply of < 1, ABS stock > 0

5 Session info

## R version 3.6.1 (2019-07-05)
## Platform: x86_64-w64-mingw32/x64 (64-bit)
## Running under: Windows 10 x64 (build 18362)
## 
## Matrix products: default
## 
## locale:
## [1] LC_COLLATE=English_Australia.1252  LC_CTYPE=English_Australia.1252   
## [3] LC_MONETARY=English_Australia.1252 LC_NUMERIC=C                      
## [5] LC_TIME=English_Australia.1252    
## 
## attached base packages:
## [1] stats     graphics  grDevices utils     datasets  methods   base     
## 
## other attached packages:
##  [1] naniar_0.4.2       kableExtra_1.1.0   sjmisc_2.8.2       summarytools_0.9.4
##  [5] skimr_1.0.7        ggridges_0.5.1     tmaptools_2.0-2    tmap_2.3-1        
##  [9] sf_0.8-0           imputeTS_3.0       janitor_1.2.0      readxl_1.3.1      
## [13] scales_1.1.0       lubridate_1.7.4    magrittr_1.5       forcats_0.4.0     
## [17] stringr_1.4.0      dplyr_0.8.3        purrr_0.3.3        readr_1.3.1       
## [21] tidyr_1.0.0        tibble_2.1.3       ggplot2_3.2.1      tidyverse_1.2.1   
## [25] pacman_0.5.1      
## 
## loaded via a namespace (and not attached):
##   [1] colorspace_1.4-1        pryr_0.1.4              ellipsis_0.3.0         
##   [4] class_7.3-15            visdat_0.5.3            sjlabelled_1.1.1       
##   [7] leaflet_2.0.3           rgdal_1.4-7             rprojroot_1.3-2        
##  [10] snakecase_0.11.0        dichromat_2.0-0         rstudioapi_0.10        
##  [13] farver_2.0.1            fansi_0.4.0             xml2_1.2.2             
##  [16] splines_3.6.1           codetools_0.2-16        knitr_1.26             
##  [19] zeallot_0.1.0           jsonlite_1.6            broom_0.5.2            
##  [22] rgeos_0.5-2             shiny_1.4.0             compiler_3.6.1         
##  [25] httr_1.4.1              backports_1.1.5         Matrix_1.2-17          
##  [28] assertthat_0.2.1        fastmap_1.0.1           lazyeval_0.2.2         
##  [31] cli_1.1.0               later_1.0.0             leaflet.providers_1.9.0
##  [34] htmltools_0.4.0         tools_3.6.1             gtable_0.3.0           
##  [37] glue_1.3.1              Rcpp_1.0.3              cellranger_1.1.0       
##  [40] fracdiff_1.4-2          raster_3.0-7            vctrs_0.2.0            
##  [43] urca_1.3-0              nlme_3.1-140            leafsync_0.1.0         
##  [46] crosstalk_1.0.0         insight_0.7.0           lmtest_0.9-37          
##  [49] timeDate_3043.102       lwgeom_0.1-7            xfun_0.11              
##  [52] rvest_0.3.5             mime_0.7                lifecycle_0.1.0        
##  [55] XML_3.98-1.20           zoo_1.8-6               hms_0.5.2              
##  [58] promises_1.1.0          parallel_3.6.1          RColorBrewer_1.1-2     
##  [61] yaml_2.2.0              quantmod_0.4-15         curl_4.2               
##  [64] pander_0.6.3            stringi_1.4.3           highr_0.8              
##  [67] tseries_0.10-47         e1071_1.7-2             checkmate_1.9.4        
##  [70] TTR_0.23-5              rlang_0.4.1             pkgconfig_2.0.3        
##  [73] bitops_1.0-6            matrixStats_0.55.0      evaluate_0.14          
##  [76] lattice_0.20-38         labeling_0.3            rapportools_1.0        
##  [79] stinepack_1.4           htmlwidgets_1.5.1       tidyselect_0.2.5       
##  [82] plyr_1.8.4              R6_2.4.1                magick_2.2             
##  [85] generics_0.0.2          DBI_1.0.0               mgcv_1.8-28            
##  [88] pillar_1.4.2            haven_2.2.0             withr_2.1.2            
##  [91] units_0.6-5             xts_0.11-2              RCurl_1.95-4.12        
##  [94] sp_1.3-2                nnet_7.3-12             modelr_0.1.5           
##  [97] crayon_1.3.4            utf8_1.1.4              KernSmooth_2.23-15     
## [100] rmarkdown_1.17          grid_3.6.1              webshot_0.5.1          
## [103] forecast_8.9            digest_0.6.22           classInt_0.4-2         
## [106] xtable_1.8-4            httpuv_1.5.2            munsell_0.5.0          
## [109] viridisLite_0.3.0       tcltk_3.6.1             quadprog_1.5-7